Football Outsiders
Innovative Statistics, Intelligent Analysis

Super Bowl XLVI Preview

by Aaron Schatz

Well, here we are again, stuck with a predicament that has become all too common. This is the ninth year that Football Outsiders has covered the NFL, and the fourth time in those nine years that a team has made the Super Bowl after following an inconsistent and overall mediocre regular season with a spectacular postseason run.

Before 2003, only one team in NFL history outscored its regular-season opponents by less than 50 points and still made a Super Bowl: The 1979 Los Angeles Rams. Since 2003, this has happened four times: the 2003 Carolina Panthers, the 2007 New York Giants, the 2008 Arizona Cardinals, and now the 2011 New York Giants, the first team to ever make the Super Bowl after being outscored in the regular season.

(Note that the 2005 Steelers and 2010 Packers, despite being sixth seeds, don't fall into this category. Those teams were very good in the regular season, but fell into a wild card spot because of bad luck in close games and/or Tommy Maddox being Tommy Maddox.)

Each of these years, in our FO Super Bowl preview, we've been stuck with the problem of figuring out just which games we should use in analyzing trends and matchups. Do we use the entire season, even when these teams were struggling? Or do we only use the numbers from the late-season and postseason surges, even though that presents us with a much smaller sample size?

In general, our research at Football Outsiders has shown us that it is better to take a longer-term view when trying to figure out how well a team will play in the future. That's why even our weighted DVOA metric, which is designed to emphasize how teams have been playing in the recent past, still measures the last eight weeks of games at nearly-full strength and the last 12 weeks of games with at least two-thirds strength. However, we can't help but anecdotally notice that this may not be the best way to analyze these surprise Super Bowl teams. Each of these years, we wondered if the team in question would turn back into a pumpkin on Super Bowl Sunday. And in each of these years, it didn't happen. The 2003 Panthers and 2008 Cardinals came within one drive of winning the Super Bowl, and the 2007 Giants of course actually pulled it off.

So with that history in mind, this year we're going to look at the last five games for each team more than we look at the full regular-season stats, with the caveat that those five-game stats represent a small sample size. For the Giants, that five-game stretch represents their current five-game winning streak, starting with their Week 16 victory over the Jets, then the must-win Week 17 game against Dallas, and then their three postseason games. For the Patriots, the last five games begin with the Week 15 win where they came back to stomp the Broncos after Denver ran for a zillion yards in the first quarter, and also includes a win over Buffalo, a too-close-for-comfort win over Miami (New England's lowest single-game DVOA in a win this year), and two playoff wins. We've made sure the game charting is complete for all ten of these games. As we noted last week on the site, the Patriots and Giants have almost exactly the same total DVOA when we compare these five-game periods.

We'll also look specifically at some stats that show how these teams played each other the last time they met, when New York won 24-20 in Foxboro back in Week 9. That game is colored yellow on the week-to-week DVOA graphs so it will stand out. However, that Week 9 game may not be the best indicator of how these teams will play on Sunday. The first Giants-Pats game was a huge aberration for both teams. Somehow, two of the strongest offenses and worst defenses in the league came into that game and went scoreless through the first half. In fact, most of the offense in that game came in the fourth quarter. The Patriots' offense had -24.8% DVOA through the first three quarters, 75.2% DVOA in the fourth quarter. The Giants' offense had -35.2% DVOA through the first three quarters, 60.9% DVOA in the fourth quarter. As a result, the yellow dot on the graphs below comes out as one of the worst offensive games for both teams, and one of the best defensive games.

For those who may be unfamiliar with the Football Outsiders stats, they are explained at the bottom of the page. Scroll down or click this link. All stats represent regular season only, unless noted. Game charting data is still incomplete, but represents most of the regular season. The Super Bowl preview includes two "week-to-week" charts for each team: one for offense, one for defense. Because defensive DVOA is opposite of offensive DVOA, the defensive charts are flipped upside-down -- the higher dots still represent better games.

Giants on Offense
DVOA 15.6% (7) 17.9% (30)
LAST 5 GAMES 26.2% (5) 12.5% (24)
PASS 36.3% (4) 24.1% (28)
RUSH -1.3% (20) 8.8% (28)
RED ZONE 1.6% (12) -15.9% (8)
Patriots on Offense
DVOA 36.8% (3) 6.9% (20)
LAST 5 GAMES 41.7% (1) -12.7% (2)
PASS 60.7% (2) 14.0% (21)
RUSH 17.1% (4) -2.0% (19)
RED ZONE 33.0% (3) 13.8% (27)
Special Teams
DVOA 0.3% (16) 3.7% (5)
LAST 5 GAMES -2.0% (27) 8.4% (1)
NYG kickoff 4.5 (8) -3.7 (25)
NE kickoff -3.4 (22) 8.7 (3)
NYG punts 10.9 (5) -0.2 (18)
NE punts -8.4 (29) 15.2 (2)
FG/XP -1.5 (19) 1.5 (14)

All readers can click here for in-game discussion on our message boards. If you have FO Premium, you can click here to see all the matchup of DVOA splits for this game.


The Patriots' defensive ratings are a good example of why we run into trouble looking at improvement over a small number of games. Look at the Patriots' rating for the season and their rating for the last five games, and it doesn't look like there is much improvement. But that's mostly because of all the running success that Denver had in Week 15. If we only look at the last four games for the Patriots, we see a lot of improvement, from 17.0% DVOA in Weeks 1-15 to 5.2% DVOA in Weeks 16-20. It's even better if we look at just three weeks, and even better than that if we look at just two weeks, because the Patriots had only four games this year with defensive DVOA below 0% and two of them were the two playoff games.

Like the Giants' defense, the Patriots' defense has been helped by getting guys back from injury. Safety Patrick Chung, the Pats' best defensive back, missed most of the second half of the season with a foot injury, and linebacker Brandon Spikes had a sprained MCL. Both players were injured in the middle of the first Giants-Patriots game, and both players returned in Week 17.

On paper, the Giants should be a horrible matchup for the Patriots defense. New England's biggest problem is poor play in the secondary. The starting cornerbacks weren't very good this year, and there the depth players are even worse. The Patriots get killed by the deep pass, and the Giants love to throw deep. And yet, as noted above, the Patriots had one of their better defensive games of the season against the Giants until Chung and Spikes got hurt and the defense collapsed in the fourth quarter. It certainly seems like an aberration that is unlikely to be repeated.

The Giants will go three-wide a lot of the time, which will bring out the Patriots' nickel and dime packages, and that means two-way player Julian Edelman. Like everyone else, I question just how successful Edelman can be trying to cover Victor Cruz or Mario Manningham. The Patriots actually shut Eli Manning down when they went to nickel and dime packages in the first game. When the Pats had five or more defensive backs on the field, Manning was 3-of-9 for 24 yards with no first downs. However, Edelman wasn't being used as a defensive back at that point in the season. In addition, this was really just a one-game trend. The rest of the season, Manning wasn't particularly poor against nickel and dime packages, and the Patriots' defense wasn't particularly better in these packages either.

(For those curious, Edelman charting stats won't be useful. We only have nine passes charted with Edelman in coverage; two of them were passes thrown away due to pass pressure, and one was a running back screen with Edelman essentially playing middle linebacker.)

Twenty-six percent of the Giants' passes during the regular season went at least 16 yards through the air, the second-highest figure in the league. That's a huge problem for the Patriots, who were the worst defense against deep passes during the regular season. They allowed 14.6 yards per pass on these deep passes (30th in the NFL), 52 percent catch rate (31st) and 69.3% DVOA (dead last). However, another stat shows you the Patriots' main strategy for preventing deep passes from completely killing them: keep everything in front of you. The Pats allowed only 3.6 average yards after catch on deep passes, fourth-lowest in the NFL. There's a lot of zone coverage, and that will also be a strategy to minimize Edelman's weakness (man coverage) and emphasize his strength (hard-hitting tackles).

In addition, defense against deep passes also happens to be a big part of the Patriots' improvement on defense in recent games. In the last five games, the Pats have allowed just 11.7 yards per pass on deep passes with 35 percent catch rate and 25.1% DVOA.

The Patriots have improved their coverage thanks to an inspired and unexpected personnel switch. Starting with the Week 17 game against Buffalo, Bill Belichick switched journeyman safety Sterling Moore and stuggling second-year cornerback Devin McCourty. For the most part, McCourty is now playing free safety while Moore is playing left cornerback. The basic idea was that McCourty played best when the ball was in front of him, so put him in a position where he would usually be facing the play. Moore had strong ball skills, so put him in a position where he could make plays on the ball, like slapping the almost-winning touchdown out of Lee Evans' hands two weeks ago or picking off two Ryan Fitzpatrick passes in that Week 17 game where he first played corner.

Now, trying to look at splits for defensive backs for only three games is a serious exercise in small sample size theatre, but there's been a massive improvement from both players since the switch. In Weeks 1-16, we have McCourty and Moore listed in coverage on 86 passes (nearly all McCourty), allowing 10.4 yards per pass with a 47 percent Success Rate. In Weeks 17-20, we have the two players listed in coverage on 19 passes (six for McCourty, 13 for Moore), allowing 2.8 yards per pass with a 79 percent Success Rate.

At the same time, Kyle Arrington over on the right (offensive left) has declined in the second half of the season. Arrington's trend is actually stronger if we look at a longer stretch -- not just the last three games or the last five games, but the entire second half of the season. Through Week 9, Arrington allowed 7.2 yards per pass with a 57 percent Success Rate. Since Week 10, including the playoffs, it has been 10.2 yards per pass with a 40 percent Success Rate.

For most of the season, Patriots opponents were so busy picking on cornerbacks with throws to their wide receivers that they didn't really bother throwing to their tight ends. Only 13 percent of passes against the Pats were to tight ends; every other defense in the league was over 18 percent. However, the Giants did use their tight end, Jake Ballard, and were very successful against the Patriots. The Giants threw to Ballard seven times in that Week 9 game, and he caught four passes for 67 yards including the game winning touchdown. This is where those injuries to Brandon Spikes and Patrick Chung played a huge role in New York's Week 9 victory. On Ballard's two biggest receptions -- a 28-yard seam route on third-and-10 and the game-winning touchdown -- he was covered by Tracy White, who plays almost entirely on special teams. White was just awful in coverage this year. We have him charted as the main defender on nine passes, with eight receptions. White has not been playing defense since Spikes returned in Week 17.

Pats' pass rush has actually improved since they lost Andre Carter for the season early in the Week 15 game against Denver. The Pats' ASR has gone from 5.9 percent through Week 14 to 10.5 percent in the last five games. And Eli Manning's sack rate has also gone up, from 4.8 percent ASR in the first 14 games of the year to 7.5 percent during the Giants' five-game winning streak.

The Patriots certainly want to get pressure on Manning, but they do not want to blitz him very much -- and they probably won't. Including the postseason, the Patriots sent just three or four pass rushers on 79 percent of pass plays this year, the seventh highest figure in the NFL. They only big-blitzed with six or more on 5.5 percent of pass plays. And while the Patriots were successful when they did choose to big blitz -- allowing just 5.1 yards per play -- Eli Manning was very strong against big blitzes this season. Including the postseason, Eli Manning had 7.2 yards per pass against three or four pass rushers, 6.8 yards per pass against five pass rushers, and an awesome 10.6 yards per pass against six or more pass rushers. In the first Giants-Pats game, the Pats only big-blitzed twice, and while one play was an incomplete pass, the other was the 10-yard touchdown to Mario Manningham that gave the Giants a 17-13 lead in the fourth quarter. The Pats rushed five in that game five times, resulting in three incompletes, an interception, and the 28-yard completion to Ballard.

There's a general impression that the Giants' running game has improved late in the season after struggling for most of the year. Based on DVOA, it really hasn't improved, but it has changed. In Weeks 1-15, the Giants' running backs averaged 3.71 yards per carry but had 48 percent success rate. Since Week 16, the Giants' running backs have averaged 4.24 yards per carry but with a 39 percent success rate. The Pats' run defense has declined if we look at the "last five games" split, but that's only because of the Week 15 Denver game. Otherwise, there's really been no change -- the Patriots are below average against the run but it isn't anywhere near as big an issue as their pass defense.

Tom Coughlin has long been known as a head coach who wants to establish his running game, but he seems to have done the wise thing and gotten away from that in 2011. Including the playoffs, the Giants ran only 43.6 percent of first downs, which ranks 29th in the NFL. The Giants were one of just three teams in the NFL who ran the ball more often on second down (44.4 percent of the time) than on first down. But the Patriots were reasonable against the run on second down, ranking 17th in DVOA. The other strength of the Pats' run defense is actually one of the big weaknesses of the Giants' running game: short yardage. The Pats ranked 11th during the regular season, allowing 58 percent conversion in short-yardage runs. The Giants converted just 53 percent, 27th in the league.

Basically, unless they are running out the clock with a lead, every time the Giants hand the ball off in this game instead of picking on the New England secondary is a win for the Patriots. It's also hard to see Bradshaw or Jacobs playing a major role in the passing game. Both backs were slightly above average as receivers, but the Patriots were very good against running backs in the passing game this season, ranking fourth in DVOA.

The Patriots' defense may have been horrible overall but it was much better in two very important situations. First of all, the Patriots improved in late and close situations, ranking 15th in the second half of close games (within a touchdown). The Giants offense actually ranks worse in "late and close" situations (13th) than it does overall (seventh). Eli Manning was awesome in the fourth quarter of close games this year (50.4% DVOA when the score was within a touchdown) but he wasn't very good in the third quarter of close games (-8.9% DVOA when the score was within a touchdown). A lot of those big Giants fourth-quarter comebacks were made possible because the Giants offense stalled and fell behind in the third quarter.

The other situation where the Patriots defense has strong numbers was the red zone, where they truly embodied the concept of a "bend but don't break" defense. As the table above shows, the Patriots turned into a top-ten defense inside the 20-yard line. You may remember that earlier this season, I wrote a piece for ESPN Insider suggesting that the concept of "red zone efficiency" was a myth -- that there's no consistency when a team either plays much better or much worse in the red zone than it plays overall. We'll do more research on this in the offseason, but if this proves to be a rule, the Patriots may prove to be the exception.

You may remember a stat we introduced back in Pro Football Prospectus 2007 called "points prevented per drive." This stat took yards allowed per drive and used that to estimate points allowed per drive. Then we looked at the difference between that number and actual points allowed per drive. The difference showed the extent to which the defense in question was a "bend but don't break defense."

The study from 2007 showed that in general, there's no year-to-year consistency where certain defenses prove to be "bend but don't break," with a couple of exceptions. One of these exceptions was the New England Patriots, and this has continued since we wrote that article five years ago. This year, the Patriots put up the best PP/Drive number in our database (drive stats currently go back to 1997). In part, this is because the Patriots allowed 37.5 yards per drive, the worst Yards/Drive figure in the database. This should project to 2.52 points per drive. In reality, the Patriots allowed 1.90 points per drive, 21st in the league.

You may be wondering if this "bend but don't break" tendency by the Patriots is real. Well, the Patriots have been above average in PP/Drive for nine straight seasons, and in every single one of Bill Belichick's seasons except for 2002. In 2010, the Patriots had the fourth-highest PP/Drive figure. The Pats own four of the top six years in PP/Drive, and six of the top 20.

Of course, if Eli Manning can hit a couple of 40-yard bombs to Cruz or Hakeem Nicks, it won't matter how stalwart the Patriots become in the red zone.


The Patriots didn't have a single game this year with an offensive DVOA below zero. But they came really close twice: in the Week 3 loss to Buffalo where Tom Brady threw four picks, and in the Week 9 game against the Giants. From the Giants' perspective, their win over New England was their best defensive performance of the year until they beat Green Bay in the Divisional round.

Why wasn't Brady as good as usual during the first three quarters of the Week 9 game? For the same reason he struggled against the Giants in Super Bowl XLII: pass pressure. It's a common perception that the Giants will not need to blitz to get a pass rush on Brady, thanks to their talented front four. During the regular season, the Giants actually blitzed just as much as any other team, but they've dialed down that tendency during the five-game winning streak.

Giants Pass Rush, 2011
Pass Rushers Weeks 1-15
Weeks 1-15
Weeks 16-20
Weeks 16-20
3-4 rushers 7.2 69% 5.1 80%
5 rushers 6.9 21% 4.0 16%
6+ rushers 3.2 9% 3.6 4%

In general, Brady kills big blitzes. Including the postseason, he's averaged 7.8 yards per pass against three or four pass rushers, 7.4 yards per pass against five pass rushers, and a ridiculous 11.6 yards per pass against six pass rushers or more. But he certainly didn't do well against the big blitz in the Week 9 Giants-Patriots game. By our count, the Giants big-blitzed five times in that game, and it resulted in four incomplete passes and a strip-sack by Michael Boley.

The Giants had to bring blitzes during the regular season because their pass rush was heavily affected by injuries, and thus not as strong as advertised. The Giants were tied for third with 48 sacks but finished only tenth in Adjusted Sack Rate, because they faced a lot of pass attempts. As we've mentioned in a couple of different articles this postseason, the ups and downs of the Giants' pass rush surprisingly matches when Osi Umenyiora was in and out of the lineup, not when Justin Tuck was in and out of the lineup. But with both of them finally playing together, the Giants' pass rush has improved during the winning streak. The Giants had 6.8 percent ASR through Week 15, and 7.8 percent ASR in the last five games. Tom Brady is also facing more pressure lately, as the Pats' offensive ASR has gone from 4.7 percent through Week 14 to 5.9 percent in the last five games. However, it will be tough for the Giants to get to Tom Brady on an early down to create a third-and-long situation. It's more likely that they take down Brady on third down to prevent a conversion. This is when the Giants bring out the four-DE "NASCAR" package that is so hard to block. The Pats' offense has 4.1 percent ASR on first/second down but 7.8 percent on third/fourth down. The Giants' defense has 6.6 percent ASR on first/second down and 8.3 percent ASR on third/fourth down. (Those numbers all include the postseason.)

We know the Giants' pass rush has improved during the five-game winning streak, but what about their coverage? That's improved too, and it is interesting to see where. No. 1 cornerback Corey Webster has decent charting numbers this season, and they haven't really changed in recent weeks. For the year, including the postseason, we have him at 6.9 yards allowed per pass and a 54 percent Success Rate. Antrel Rolle has had his troubles, but that hasn't changed during the winning streak either. For the year, including the postseason, we have him at 8.7 yards allowed per pass and a 51 percent Success Rate.

Nope, based on our game charting, the biggest difference in the secondary has been Aaron Ross. Ross was terrible for most of the year. Our charting numbers for Weeks 1-15 list him with 10.0 yards allowed per pass and a 42 percent Success Rate. Those are two of the worst figures in the NFL this season. Since Week 16, we've charted him as the main defender on 17 passes, allowing 4.7 yards with a 53 percent Success Rate. That includes a couple of deep passes where Ross was directly responsible for the pass defensed. I have no idea if this improvement is real or just small sample size.

The other big difference in recent games is the number of passes that were marked either "uncovered" or "hole in zone" by our charters. This could suggest that the Giants are having success playing more zone coverage, which matches the general storyline I've read of "Perry Fewell went back to basics." In Weeks 1-15, the Giants had 3.9 passes per game that were marked hole in zone or uncovered, and these plays gained 10.9 yards per pass. Since Week 16, it is 8.8 passes per game gaining 7.3 yards per pass.

If we want to look at the Giants' coverage from a different angle, using the "defense vs. receivers" numbers instead of game charting coverage stats, it is clear the Giants have big problems with "other receivers," where the Giants rank 30th in DVOA. However, this may not be a big problem against the Patriots. The Pats threw only eight percent of passes to players in this category, which ranked 31st in the league. Often, the players who fall into the category of "other receivers" are slot receivers, which would seem to suggest that the Giants will have particular trouble with Wes Welker. You don't need DVOA stats to figure that the Giants have trouble covering Welker. In Week 9, Welker caught nine of 10 passes for 136 yards. However, despite Welker's big day, the Giants' problem with "other receivers" doesn't seem to be an issue with slot guys. The players who had big days against the Giants are players who are more downfield threats: Danario Alexander, Laurent Robinson, Devery Henderson, Doug Baldwin.

Of course, the most important receiver for the Patriots isn't Wes Welker anymore. It's tight end Rob Gronkowski. We don't know how well Gronkowski will be able to play due to his high ankle sprain, but there's very little question that he's going to play, and Aaron Hernandez will be out there with him. The Giants were pretty good against passes to tight ends this year, ranking 12th in DVOA. They were an average defense when opponents used two tight ends in the formation, which is a little better than they ranked overall during the regular season.

Gronkowski was a huge part of the Patriots' game plan in the Week 9 game, with a season-high 15 pass targets. He caught eight of those passes for 101 yards and a touchdown. As Vince Verhei pointed out in a column for ESPN Insider this week, the Giants used different defenders against Gronkowski depending on the depth of his pattern. Outside linebacker Mathias Kiwanuka had him on the short routes, middle linebacker Boley covered him on the midrange passes, and safety Deon Grant picked him up when he went deep, with Ross and strong safety Kenny Phillips making cameo appearances here and there. Grant was the most effective defender against Gronkowski, covering him on five throws which all ended up incomplete. Grant is known as more of a hard hitter than he is a coverage safety, but if he can repeat that performance again, it will go a long way towards delivering the game for New York.

Although the screen pass is often a good strategy against a team with a strong pass rush, the Giants allowed just 5.2 yards per pass on running back screens, eighth in the NFL. They allowed 6.0 yards per pass on wide receiver or tight end screens, which ranked 18th. The Patriots run the wide receiver screen a lot more often than they run the running back screen. We only have them charted with 11 running back screens, although they gained an average of 11.4 yards. They gained 7.0 yards per pass on wide receiver/tight end screens, which ranked fifth in the NFL.

Most readers know that Football Outsiders generally advocates that teams should pass much more than they run, but this game might be an exception for the Patriots. The Giants may have the reputation as a "running team," but it is the Patriots who actually have the efficient running game. It is efficient and consistent at breaking out 4-to-6 yard runs, but not much more than that; the Patriots ranked 30th in Open Field Yards per carry (yards gained more than 10 yards past the line of scrimmage).

Despite all the improvement in the Giants' pass defense over the last five weeks, there's been no improvement in their run defense. It has basically been average all season; the Giants have allowed slightly fewer yards per carry since Week 16, but against a weak schedule of opposing backs. The Giants are weakest against runs around left end and runs up the middle (26th in ALY in both). And the Giants are particularly susceptible to runs out of spread shotgun sets, which is how the Patriots are often running the ball. Including the playoffs, the Giants allowed 6.59 yards per carry on shotgun runs, 27th in the NFL. This is the one place where the run defense has improved in the last few weeks, though; the Giants have allowed just 5.22 yards per carry on shotgun runs since Week 16.

Running the ball won't just gain effective, consistent yardage. It will also set up the Patriots' great play-action passing game. Only 17 percent of Patriots' passes during the regular season were deep passes over 15 yards through the air, which ranked 29th in the NFL. But when the Patriots go deep they usually do it with a play fake first, and they confuse their opponents. The Patriots this season gained 10.9 yards per pass with play action, compared to 7.4 yards per pass otherwise. This was the highest figure of yards per pass with play action in the league, and the third-highest gap between play action and non-play action passes. During the regular season, the Giants had trouble with play action, allowing 7.9 yards per pass with play action compared to 6.0 yards per pass otherwise. During the five-game win streak, the Giants have still given up more yardage to play-action plays, but with a gap that matches the league average (5.9 yards with play-action, 4.7 yards otherwise).


If we look at just the stats from the last five weeks, the Patriots have been far better than the Giants on special teams. But that's a bit silly. The sample size problem created by looking at just five games on offense or defense is pittance compared to the sample size problem created by looking at just five games of special teams. Really, the Patriots have just a small field position advantage from special teams. Specifically, both teams get a field position advantage whenever they kick off or punt, but the Patriots' advantage is a little bit bigger. The Patriots are outstanding in both net kickoff and net punt value, and the Giants are pretty good. Neither team has a record of strong returns this year. Neither team has a kickoff return longer than 40 yards, and a 72-yard punt return touchdown by Edelman back in Week 11 is the only punt return by either team longer than 25 yards.


This is a strange game to pick. The math clearly favors the Patriots. Our various "pick this game" methods generally look at things longer-term, using weighted DVOA, and the Patriots come out significantly better than the Giants if you look at any period longer than just the last five weeks. That's why, for example, the Patriots are favored in the playoff odds report. Las Vegas bookmakers, likely using methods similar to ours, installed the Patriots as three-point favorites. It's moved to -2.5 at some sports books, but no further. And yet, the Giants are ridiculously overconfident, the majority of the betting public (around 65 percent) has put money on the Giants, and nearly every prognosticator out there is picking the Giants to win this game.

It's a close call, but I'm afraid I lean that way as well.

Obviously, football games can turn on a lot of different variables. There are 90 players on the field, the ball might take a strange bounce on any given play, a player will come within inches of making a great catch or not. Still, I think there are two driving issues that will do the most to determine who wins this game:

1) How healthy is Rob Gronkowksi, and

2) How real are the Giants' and Patriots' recent defensive improvements?

With Gronkowski not practicing all week, it is hard to believe the answer to the first is 100 percent. He's going to play, no doubt about it, but we don't know at what level of effectiveness. As far as the second question, the Giants' recent defensive improvement is easier to believe in than New England's. Both teams have gotten healthier on defense, but the Giants have improved more, their improvement has a larger sample size, and their raw defensive talent is just plain better. "The Giants have a ferocious pass rush" just seems like a more logical statement than "it turns out Sterling Moore is a really good cornerback and nobody ever noticed." If Gronkowski is a little hobbled, and Brady is a little bit pressured, the Giants' offense should have enough weapons to attack the Patriots' porous secondary and come away with a close victory. Math may favor the Patriots, but matchups and circumstance favor Big Blue.


DVOA (Defense-adjusted Value Over Average) breaks down each play of the season and compares it to the NFL average based on situation and opponent. You'll find it explained further here. Since DVOA measures ability to score, a negative DVOA indicates a better defense and worse offense, and a positive DVOA indicates a better offense and worse defense.

SPECIAL TEAMS numbers are different; they represent value in points of extra field position gained compared to NFL average. Field goal rating represents points scored compared to average kicker at same distances. All special teams numbers are adjusted by weather and altitude; the total is then translated into DVOA so it can be compared to offense and defense. Those numbers are explained here.

Each team is listed with DVOA for offense and defense, total along with rush, pass, and red zone, and rank among the 32 teams in parentheses. (If the DVOA values are difficult to understand, it is easy to just look at the ranks.) Each team also gets a chart showing their performance this year, game-by-game, according to total DVOA. In addition to a line showing each game, another line shows the team's trend for the season, using a rolling average of the last five games. There are separate charts for offense and defense for each team. The yellow dot in Week 9 represents the regular-season game between the Patriots and Giants.


218 comments, Last at 23 Apr 2013, 10:56pm

1 Re: Super Bowl XLVI Preview

As a Patriots fan,unfortunately I feel the same way--a bit more confident than 2001, but not by much.

67 Re: Super Bowl XLVI Preview

New England will win this game, purely because I want the Giants to win. So far this playoffs, every team I have rooted for has lost. I guess I'm the Schleprock of football viewers.

131 Re: Super Bowl XLVI Preview

I like the gameplan of running the ball a lot for NE. If the Pats are balanced in this game, I actually think the Pats enjoy a significant LOS advantage in this game. Of course, the Pats are going to need to prevent their patchwork secondary from being exploited, but I think the numbers are right here. NYG have a very poor rushing attack, their o-line has some obvious and significant weaknesses, and their d-line (and defense more generally) is best geared to stop vertical passing attacks.

If Ridley is out of the doghouse, I think he could have a nice game. Assuming he doesn't fumble.

164 Re: Super Bowl XLVI Preview

Exactly. They also ran well against the Ravens last week, which is much better than the Giants in preventing the run. Keep Brady on the field and give less time for the secondary to be exploited, and I think the Pats may pull out a win.

98 Re: Super Bowl XLVI Preview

Yes, I agree. I definitely feel slightly more confident playing a team that was significantly worse than us in the regular season as opposed to a team that was enormously better.

Give me a break.

128 Re: Super Bowl XLVI Preview

I will have my jersey on and be in my assigned position in front of the TV, ready to perform my assigned function of yelling "You yellow bellied coward!" at the coach every time a punter comes onto the field.

Does that count?

2 Re: Super Bowl XLVI Preview

Largely agree, especially if Gronk is hobbled enough to allow the Giants to take away Hernandez without sacrificing much. The Pats haven't faced a terrific qb in a long time, and the last game is a bad time to get reacquainted with the species.

HOWEVER, the Giants are merely serviceable, at best, on the o-line, and a team like that is always in danger of a brutal clubbing, and a defensive front with a Wilfork playing at a very high level has a decent chance of administering such a clubbing.

The game is a toss-up, it seems to me, and a damned interesting one. Force me to pick, and I guess I say New York, but without any confidence to speak of.

7 Re: Super Bowl XLVI Preview

The last 'terrific' QB the Pats faced was Eli.

Sanchez, Palko, Vince Young, Orlovsky, Grossman, Tebow, Matt Moore, Fitzpatrick, Tebow, Flacco. This is a step up. It isn't facing Aaron Rodgers or Drew Brees, but I would say Manning and those receivers are significantly better than any of those previous ten offenses.

The o-line doesn't bother me as much because Eli Manning is good at getting the ball out quick. He's not as good as Peyton in that area, but he's had a bad o-line all year. The 49ers killed him, but the 49ers pass rush, in my eyes, is better than NE's no matter how much the Pats pass rush has improved over the year.

I have no idea where to lean in this game. When Championship Sunday ended two weeks ago, I was fairly certain that the Giants would win. Ever since my confidence in that has been eroding. This is Tom Brady and the Pats. Their defense is very good in the red zone. Eli will still throw a couple up for grabs. I still think the Giants will win, but not by too much (that said, I can see a Giants blowout as more likely than a Pats blowout - if the Giants pass rush plays its best, and the offense does what it does and the Pats defense is 'what we thought it was' the Giants can win comfortably).

As for Gronk. I have a bad memory about this. If we forget about the ridiculous hype that the "is Maurkice Pouncey going to play" storyline got last year, the last big injury going into a Super Bowl was Freeney two years ago. I believe Freeney that time also had an ankle sprain. He played, but was less than 100%. He made plays in the 1st half (sacked Brees with one-hand), but by the 2nd half was just done. I could see a similar thing for Gronk, where he plays, but just is nowhere near the same. I think Hernandez was a little gimpy in the November game, and although Gronk had a good day, the Giants played the Pats quite well defensively.

9 Re: Super Bowl XLVI Preview

The Giants really didn't play that well defensively in the November game. Brady just had an abominable first half partly due to an elbow injury and partly because he was seeing ghosts. Once he settled down in the second half, NY had no answer.

Hernandez was also far more than simply "a little gimpy". He wasn't up to to speed for another month or so.

All that said, I agree that a NY blowout is more likely than a NE blowout, and that NE hasn't faced a passing unit of this caliber for some time.

13 Re: Super Bowl XLVI Preview

I don't think the Giants played tremendous defense, but they played definitely well enough to win. Brady was seeing ghosts for a reason. The Giants didn't hit him, but definitely he felt the defense coming in. I also thought they covered well. Brady did throw two interceptions that I both thought were bad throws.

17 Re: Super Bowl XLVI Preview

Yeah, there was a reason he was seeing ghosts, but it really didn't have much to do with pressure applied during that game. JPP was right that they were in Tommy's head.

The pics were terrible throws. The one to Gronk, Brady got cute and didn't even bother to read the coverage after looking off the safety. Gronk was there, but only if the throw was lofted more.

On the other one, Brady saw Deion flash behind the LBs and just tossed it up without even checking if anyone was sitting in the zone.

In both cases, the coverage was good, but the QB was mostly responsible for the turnover.

28 Re: Super Bowl XLVI Preview

I think it was the pick to Deion where from the angle, it did originally look like Deion was wide open, but Brady waited for some reason for him to continue his crossing route across the field, at which point the players in the zones converged and picked it off. That was just a bad Brady throw in a pretty off Brady game.

99 Re: Super Bowl XLVI Preview

The pics were terrible throws. The one to Gronk, Brady got cute and didn't even bother to read the coverage after looking off the safety. Gronk was there, but only if the throw was lofted more.

If that's the case, he was doing the same things in the Baltimore game. Brady occasionally gets the yips.

120 Re: Super Bowl XLVI Preview

He does.

That being said, the picks two weeks ago weren't the yips. #1 was laziness and #2 was bullheadedness. If he gets the same protection vs. the Giants that he did vs. the Ravens, he's going to put up some big, big numbers.

123 Re: Super Bowl XLVI Preview

Why do you assume he won't be lazy or bullheaded again?

First, I doubt he gets that protection again. The Ravens rush got worse as the season wore on, while the Giants rush has gotten better.

Brady can be bullheaded. He does make bad throws. Even in his good games (the pick against Denver). He's not perfect, and just because he had a bad game doesn't mean he is going to have a good to great game next. In my opinion, Brady's bad games usually come in bunches (like the middle of this year with the DAL/PIT/NYG stretch or last year with the BAL/SD/MIN/CLE stretch). This is not going to be an easy game.

130 Re: Super Bowl XLVI Preview

The "bad games in bunches" is a very recent phenomenon. Until the last year or so, Brady had a well documented stretch of putting together impressive rebound performances. I also think it needs to be noted that he was playing with a significant elbow issue during the 2011 games you mentioned, severe enough to have Ben Roethlisburger running for the nearest microphone.

The reason why he might not have the same issues as the AFCCG is that Baltimore is dramatically better than NY in the back seven. They actually did a decent job of sticking with receivers even when the rush didn't get there, and it is hard to see NY doing as well. I agree that there were signs that Baltimore would struggle to get to Brady, whereas NY has been playing their best ball of late. That said, it has been a while since NY has faced a good offensive line.

138 Re: Super Bowl XLVI Preview

Atlanta did rank #7 in ASR for an o-line, and DAL ranked 13th, so they haven't played any great o-lines, but they haven't feasted on horrible ones either (Jets 17th, San Fran 23rd and GB 25th were the other three games).

The Pats are 8th. I don't expect a repeat of Super Bowl XLII, but I wouldn't be surprised if the Giants d-line gets to Brady quite a bit.

141 Re: Super Bowl XLVI Preview

Speaking of SB 42, watching the highlights it really struck me how much faster Tuck looked only 4 years ago. His explosiveness looked downright JPP-esque, whereas now he relies much more on his excellent technique to get in the backfield. Every athlete loses a half-step from his early 20s to his late 20s, but I think overrating of the NYG pass rush stems from the fact that people think he is the guy he was back then, when the eye test says that he's not. On the other hand, 30 year old Osi looks as fast as ever despite similar injury history. If the Giants let Osi go this offseason, they will miss him a lot.

26 Re: Super Bowl XLVI Preview

Drew Brees plays behind the best offensive line in the league, with at least nine games in a domed stadium. Rodgers was great, obviously, but he came out of the chutes with a receiving corps that was primed to produce. Manning plays behind a mediocre line, in a stadium not conducive to passing, and with a receiving corps that took a good chunk of the season to gel. No, I'm not saying I would take Manning over Rodgers (I might take him over Brees if I'm playing in the Meadowlands behind a significantly lesser o-line), but I really think we overestimate the degree to which the production metrics mirror individual performance. Manning, as you note, is a step up from what the Pats have seen since they last played the Giants, and I'd argue a huge step up from the likes of Tebow, Palko, Orlovsky, Young, Grossman, or Moore. Just Manning's ability to get the Giants into the right play, against a secondary with significant vulnerabilities, makes a big difference.

I think Brady is going to have some good drives, and if Gronk is effective the Pats will run efficiently, taking some starch out of the rush. The more I think about the Giants, the less I like their offensive line, and an offensive line having a bad day is good formula for defeat. The Pats defensive front isn't wonderful, but I could see them thumping the Giants o-linemen. I'll still lean Giants, but I think I see a clear path for me to be wrong.

29 Re: Super Bowl XLVI Preview

I agree completely. I think the difference between Eli and Brees/Rodgers is much closer than conventional stats or advanced stats say.

This is, at least for 2011, the best QB the Pats have faced all year (of course, they've already faced him). The Pats weren't great against two of the other three top-10 QBs they faced (Rivers, Roethlisberger) - the other was Romo, who they played well, but is probably a borderline top-10 for most people. Rivers and SD pretty much went up and down the field with only the usual Norv-Chargers stuff getting in the way (failed 4th and goal from the 1, Wilfork picks off a pass, Mathews fumbles on a normal run play with the Chargers down 7).

49 Re: Super Bowl XLVI Preview

Brees played a whopping 11 games in a dome this year...the eight home games, and away games at ATL, MIN, and STL. And I agree...with the exception of maybe HOU, he has the league's best o-line.

Not that he isn't awesome, but those are real advantages.

51 Re: Super Bowl XLVI Preview

Brees is also more dependent on good o-line play, due to his height. He's great, no doubt about it, but of the guys constantly talked about as being among the best, he's the one guy who would be most likely to see his production suffer a steep decline if his protection was not good.

52 Re: Super Bowl XLVI Preview

I have to disagree. Brees is short, but he also (1) reads the defense and makes decisions really, really quickly, (2) moves extremely well, and (3) throws accurately on the move. His production would decline as any QB would, but I think he'd adjust.

I think Flacco and Sanchez are the two QBs most dependent on good OL play. Flacco in particular doesn't seem to have very good pocket awareness, and takes a very long time to go through his reads. I think Baltimore's OL is a lot better than they're widely given credit for.

55 Re: Super Bowl XLVI Preview

I agree with Will. I watched Brees very closely this postseason and his dependency on good protection up the middle is dramatic. This is because he tended to walk up into his throws more than simply stepping up. Because of this, to be precise he needed three full yards open in front of him, whereas guys like Manning and Brady only need a yard and a half to stride into.

I don't watch Drew all the time, so these two games might not have been representative of his standard style of play, but it was notable against Detroit and SF.

178 Re: Super Bowl XLVI Preview

Manning plays behind a mediocre line, in a stadium not conducive to passing, and with a receiving corps that took a good chunk of the season to gel.

He also has poor accuracy and consistency relative to the other QBs being discussed.

Why has everybody forgotten this?

161 Re: Super Bowl XLVI Preview

I think it's interesting that no one's looked at the defensive DVOA splits for the games against good QBs (to me: Romo, Roethlisburger, Rivers, Flacco, Manning). The graph suggests they played 5 of their top 7 defensive games against those QBs. Weirdly the worst games seem to have come against the unstoppable machines of Miami, Indy, Buffalo and the Jets. Obvs that's taking into account the run games too and it's DVAO, not VOA, but still - some evidence for the idea of 'playing up to the level of your opponent' ?

179 Re: Super Bowl XLVI Preview

I would throw out the stats from the Indy game. The Pats had a 31-3 lead going into the fourth quarter and got lazy.

The other games you mention are divisional games. It happens.

199 Re: Super Bowl XLVI Preview

I think a simpler point is this: who the Patriots played didn't really have that much impact on the passing production they allowed. Against a weak team and negative adjustments, they looked atrocious. Against a good team and positive adjustments, they looked average to mediocre.

There's good reason to think Eli is going to have an average "QB playing the Patriots" day, not something substantially better.

3 Re: Super Bowl XLVI Preview

I have a hard time seeing the Pats rush getting to Eli. All of my guesstimations are based on that fact. Therefore, I think if the Giants hammer Brady like we're all kind of expecting, they win. If the Pats' o-line holds up, we're in for a shoot-out, and I think the Patriots come out on top. I have absolutely no idea how to pick this one, but I'm leaning New England.

(Full disclosure: I'm a Lions fan. I have no idea if this says anything about inherent biases or not.)

"When you absolutely don't know what to do any more, then it's time to panic." - Johann van der Wiel

6 Re: Super Bowl XLVI Preview

I have a hard time seeing it mostly because Eli can move around the pocket and get rid of the ball, and not because the Giants OL is likely to stop Wilfork, Love, or Anderson unless they trip over David Diehl's supine body. For all the 2007 comparisons, I think we're overlooking the very real possibility that the Pats DL might dominate the LOS cover up a weak secondary the way the Giants did four years ago.

I'm obviously rooting for the Giants, and think they match up well with the Pats, but I can't bring myself to call them favored in any way. I have a sinking feeling in my gut that (1) the Giants, who re-defined the term 'mental lapse' throughout the regular season, are overconfident instead of hungry, and (2) Belichick and Brady have images of David Tyree & Justin Tuck seared into their brains over the last four years, and are filled with sheer hatred an malice.

That's not scientific, I know, but just a deep, deep sense of unease in my gut.

21 Re: Super Bowl XLVI Preview

I share your concern about the LOS matchup. I think the game is a coin flip, given Gronkowski having some effectiveness. However, I don't think there's any danger of the Giants being in the wrong mindset/being overconfident/having a mental lapse. Their mystifyingly bad games always come against the Redskins and Seahawks of the world, not in big games against good teams. As for the revenge angle, it's the Superbowl. Everybody's motivated. If anything, it could actually be an advantage to the Giants if Brady comes out tight and forcing things, because that's when he plays badly.

32 Re: Super Bowl XLVI Preview

The past two years the Giants played the eventual world champs and stunk up the joint in crucial-for-them games, although the Saints game was only crucial in retrospect, because it revealed how awful their CCsecondary was

137 Re: Super Bowl XLVI Preview

The 2009 Saints were just that much better than the 2009 Giants, even though the Giants had a good record going into that game. The Giants came into that game after feasting on the AFC West, and just weren't that good. I also wouldn't call that game particularly crucial, as it came in week 6 and the Giants were on top of their division at the time. The 2010 Packers loss was obviously much bigger, being basically a week 16 playoff game. Coming after the Miracle at the Meadowlands II, though, I don't think it applies to the current Giants. They were mentally shattered after that game in a way that is extremely rare.

11 Re: Super Bowl XLVI Preview

For all of the talk of NY's dominant pass rush, NE has comparable playoff sacks/pressures despite playing one less game, 9/18 vs. 8/15. It isn't as if NE's are skewed by game circumstances, either, since NY has had similar big lead advantages.

This isn't the quicksand pass rush the Patriots have employed the past three years or so, these guys are actually pretty good. If NY was bringing the 2007 OL into this game, I'd write it off as having no chance, but there are some attractive matchups on thst side of the ball, certainly as many as NY's DL has against NE's blockers.

15 Re: Super Bowl XLVI Preview

I agree that the Pats DL has been playing well, Wilfork in particular, but if there's a skew in the numbers, it's that those two games were against QBs renowned for their lack of pocket awareness. I'm not sure where your numbers are coming from, though; Pro Football Focus had the Alex Smith being pressured 24 times, and Rodgers 16 times in their respective games, which roughly jibes with my memories of those games.

19 Re: Super Bowl XLVI Preview

I took them off ESPN, though you are write that I mislabled them. It should be sacks/hits, not pressures.

As for the caliber of QB, you might be right, but SF's pass blocking was lousy most of the year, and GB's OL wasn't full strength. Some competing factors, though I'm not sure how to weigh each.

18 Re: Super Bowl XLVI Preview

It is hard to say that the Giants had similar big lead advantages. They were up 10 or 7 points for most of the second half of the Packers game (until the Kuhn fumble and then TD by Manningham). They didn't have more than a one-score lead in the 49ers game. I think the Pats pass rush is real though, but I think they've had some fortunate situations the last two weeks (Tebow is awful at getting rid of the ball).

4 Re: Super Bowl XLVI Preview

Personally, as an advocate of advanced statistical analysis and its strength over individual assessment, I'm rather disappointed in Aaron's final paragraph.

Schatz decides to arbitrarily assign + value to certain things over others (like the improvement of the Giants pass rush vs. that of the New England D) and comes off as just not trusting the numbers.

Of course statistical models are never made to pick any one individual game. But we know over the course of thousands of games, they will almost always perform better than any one human assigning arbitrary values to whatever part of the game they feel will be important.

Baseball statisticians have proven (in large part thanks to larger sample sizes than football) that there are no true "Mr. Octobers". That any player, given enough time, is no better during the playoffs than they are during the regular season. We know by looking at frequency distributions of 5 game streaks across the season for all teams that there's no such thing as "true" hot or cold streaks, and that the best way to evaluate this game (like any. other. game.) is using weighted stats that the Weighted DVOA model provides.

But Aaron just doesn't trust the numbers this time. There must be something wrong with the model if there are 4 odd results in the past 9 Superbowls.

Sometimes when you flip a coin 10 times it comes up heads 8. That doesn't mean we should re-evaluate the accuracy of our coin flipping predictive model. No matter what "matchups" or "circumstance" we choose to perceive.

8 Re: Super Bowl XLVI Preview

You would rightfully have a much higher confidence in your statistical model for coin flipping than for NFL game outcomes before running any experiments at all. Ten samples isn't a lot to go on, but when that's all you have that's what you have to use.